We plan to evaluate the relationship between various intermediate outcome measures and mortality from breast cancer in a screening setting. Group Health Cooperative of Puget Sound serves as a unique site for a cohort approach to the problem, since it serves a large defined population of women 40 years of age and older (68,796), has an established SEER registry to permit identification of all cases and deaths, an active risk-based breast cancer screening program which provides risk factor information and assures large numbers of screened and unscreened women, and a unified medical record system permitting retrieval of screening history and detailed surgical and pathological examination information for all members of the defined population. The analytic approach (hybrid case-cohort design) is a modification of the case-cohort approach of Prentice which is typically more efficient statistically than case-control designs and much less costly than a standard cohort design. Important sources of bias are carefully considered in our approach.